F430 Model Chemistry. Mechanistic Investigation of the Reduction

Aug 7, 1992 - syringe is too low to permit adequate determination of the mass balance ... and the head-space, sample loop, and injector dead-space vol...
7 downloads 0 Views 1MB Size
Download PDF
4991

Inorg. Chem. 1992, 31, 4991-5000

F430 Model Chemistry. Mechanistic Investigation of the Reduction, Coupling, and Dehydrohalogenation of Alkyl Halides by the Nickel(I) OctaethylisobacteriochlorinAnion Goutam K. Lahiri, Leonard J. Schussel, and Alan M. Stolzenberg' Department of Chemistry, West Virginia University, Morgantown, West Virginia 26506-6045

Received August 7, 1992 The reactions of NiI(0EiBC)- with alkyl halides andp-toluenesulfonates and with hydrogen atom and proton donors were investigated in a wide range of solvents. The ease of reduction of Ni"(0EiBC) by sodium amalgam and the stability of the Ni'(0EiBC)- generated was greater in more polar solvents. NiI(0EiBC)- was not obtained in solvents with weakly acidic protons like N-methylformamide. Ni*(OEiBC)- is oxidized by weak acids like water, alcohols, and thiols to afford HZin a reaction with an apparent stoichiometryof 2:l Ni(I):H2. Reaction with CHjX, X = I and OTs, afforded neutral NiII(0EiBC) in quantitative yield and CH4 and CzH6. Trace quantities of C2H4 were also observed. The mass balance of the reaction was essentially complete in THF, but the yield of volatile products was less than expected in other solvents. Reaction with CzHsX afforded hydrocarbons C2H6, QH4, and C4HI0and trace quantities of C4Hs. The relative yields of products in both reactions was strongly dependent on the solvent. Coupling (CzH6 from CH3X and C4HI0from C ~ H S Xand ) elimination products (C2H4 from C2HsX) increased in more polar solvents. More coupling was observed when X = I than OTs. Isotopic labeling experiments established that the source of the fourth hydrogen in CH4 is solvent and a pool of protons that is probably residual water. Simultaneous addition of alkyl halides and weak acids led to competitive production of H2 and alkanes. The proportion of coupled alkyl products decreased greatly in the presence of the weak acid. These observations and other evidence rule out free-radical processes as all but minor contributors to the mechanism and suggest that the reduction, coupling, and dehydrohalogenationof alkyl halides occur at the nickel atom. A mechanism that involves transient alkyl-Ni(0EiBC) and H-Ni(0EiBC) intermediates is proposed. It explains adequately the distributions of products and the production of the trace products and suggests that the unusual structural preferences of Ni(1) could be of functional significance. The apparent failure of NP(OEiBC)-and of Ni'F430M to react with thioethers, including methyl coenzyme M, is discussed in light of the mechanism.

The reactions of the nickel(1) octaethylisobacteriochlorinanion, NiI(OEiBC)-,l with alkyl halides and other electrophiles model many features of the reactions of F430.2 F430 is the nickel

\COzH

-0

Ni(0EiBC)

F430

hydrocorphinoid prosthetic group of the enzyme methyl coenzyme M reductase,3 which in the final step in methanogenesis catalyzes the reductive cleavage of the thioether methyl coenzyme M in the presence of the thiol cofactor (7-mercaptoheptanoy1)threonine phosphate (HS-HTP) to afford methane and a mixed disulfide, eq 1. The enzyme may also be involved in the reductive dehalogenation of chlorinated hydrocarbons by methanogenic (1) Abbreviations: OEP, 2,3,7,8,12,13,17,18-octaethylporphyrindianion; OEC, trans-2,3-dihydro-2,3,7,8,12,13,17,18-octaethylporphyrindianion; OEiBC, mixture of ttt- and tct-2,3,7,8-tetrahydro-2,3,7,8,12,13,17,18octaethylporphyrin dianion; F430M, the pentamethyl ester of F430; TPP, 5,10,15,20-tetraphenylporphyrindianion; tmc, R,RSSor R,S,RS isomer of 1,4,8,1l-tetramethyl-1,4,8,1 I-tetraazacyclotetradecane;teta, 5,5,7,12,12,14-hexamethyl-l,4,8,11-tetraazacyclotetradecane; dmg, dimethylglyoximato. (2) Sto1zenberg.A. M.;Stershic, M.T.J. Am. Chem.Soc. 1988,110,5397. (3) (a) Ellefson, W. L.; Whitman, W. B.; Wolfe, R. S.Proc. Natl. Acad. Sci. U.S.A. 1982, 79, 3707. (b) RouvWre, P. E.; Wolfe, R.S.J . Biol. Chem. 1988, 263, 7913. (c) Friedmann, H. C.; Klein, A.; Thauer, R. K. FEMS Microbiol. Rev. 1990, 87, 339.

0020-1669/92/1331-4991$03.00/0

CH,-S-CoM

+ HS-HTP

-

CH4

+ COMS-S-HTP

(1)

b a ~ t e r i a .The ~~~ nickel atom of F430 undergoes oxidation state changes during catalysis by the intact, active enzyme.6 With the exception of F430,' nickel isobacteriochlorin complexes8~9are the only tetrapyrrolic compounds that afford isolable nickel(1) complexes when reduced. The nickel(1) forms of F430"J and of OEiBCZJareact rapidly with methyl iodide to afford significant quantities of methane. Our initial study of the reactivity of NiI(0EiBC)- revealed some important features.* The anion reacts with methyl iodide in a 2:l stoichiometry to afford neutral NiI1(OEiBC) in quantitative yield and methane in lesser yield. It also reacts with methyl p-toluenesulfonate, MeOTs, but the amount of methane produced is smaller than with methyl iodide. Ni(0EiBC) mediates the electrocatalytic reduction of alkyl halides and p-toluenesulfonates at the potential of the Ni(II)/Ni(I) couple. We analyzed the current vs potential curves of the electrocatalytic reductions to obtain relative second-order rate constants for a (4) Krone, U. E.; Laufer, K.; Thauer, R. K.; Hogenkamp, H. P. C. Biochemistry 1989, 28, 10061. ( 5 ) Fathepure, B. Z.; Boyd, S.A. FEMS Microbiol. Lett. 1988, 49, 149. (6) (a) Albracht, S.P. J.; Ankel-Fuchs, D.; van der Zwaan, J. W.; Fontjin, R. D.;Thauer, R. K. Biochim. Biophys. Acta 1986,870.50, (b) Albracht, S.P. J.; Ankel-Fuchs, D.; Bticher, R.; Ellerman, J.; Moll, J.; van der Zwaan, J. W.; Thauer, R. K. Biochim. Biophys. Acta 1988, 955, 86. (7) Jaun, B.; Pfaltz, A. J . Chem. SOC.,Chem. Commun. 1986, 1327. (8) (a) Stolzenberg, A. M.; Stershic, M. T. Inorg. Chem. 1987, 26, 3082. (b) Stolzenberg, A. M.; Stershic, M. T. J. Am. Chem. Soc. 1988, 110, 6391. (9) (a) Renner, M. W.; Furenlid, L. R.; Barkigia, K. M.; Forman, A,; Shim, H.-K.; Simpson, D. J.; Smith, K. M.;Fajer, J. J . Am. Chem. SOC.1991, 113, 6891. (b) Furenlid, L. R.; Renner, M. W.; Smith, K. M.; Fajer, J. J . Am. Chem. SOC.1990, 112, 1634. (c) Renner, M. W.; Forman, A,; Fajer, J.; Simpson, D.; Smith, K. M.; Barkigia, K. M. Biophys. J. 1988, 53, 277. (10) Jaun, B.; Pfaltz, A. J . Chem. Soc., Chem. Commun. 1988, 293.

0 1992 American Chemical Society

4992 Inorganic Chemistry, Vol. 31, NO.24, 1992 series of substrates. The reactivity trends I > Br > OTs > C1 and CH, > n-CsHg > sec-CdH9 > tert-CsHg are consistent with a two-electron nucleophilic mechanism. This is remarkable because the reduction potential of N i ( 0 E i B C ) is considerably negative of nickel tetraazamacrocycliccomplexes,yet the nickel( I) forms of the latter react via electron-transfer and/or radicalbased mechanisms.lL-13 The evidence led us to propose that a transient alkyl-Ni(0EiBC) complex is a n intermediate in the

reaction.2 Several important issues were not resolved by the initial study. The mass balance of the reaction was incomplete. The source of the fourth hydrogen in methane was not determined. Direct evidence for an alkyl-nickel intermediate was not obtained. The distribution of products appeared to depend on the choice of substrate and solvent and perhaps on other unidentifiedvariables. I t was not clear whether the course of the reaction is affected by the manner in which it is run. Reductions of substrates by NiI(0EiBC)- can be conducted in batch or in catalytic reactions. The latter can be driven either chemically or electrochemically. Finally, the observation that 1-hexene and lesser amounts of methylcyclopentane were produced upon reduction of the radical clock probe 1-bromo-5-hexene raised the possibility that radical species could play a role in the reaction mechanism following the rate-determining step. The work in this paper addresses the questions of product mass balance and solvent effects on the reactivity. The investigation focuses on electrophiles t h a t contain the two simplest alkyl groups-ethyl, which has a @-hydrogen,and methyl, which does not. Wealsoexamine thereactionsofNP(0EiBC)-withhydrogen atom and proton donors. A thorough appreciation of the factors that can affect product distributions with these simple substrates is essential if experiments with stereochemical and radical-trap probes are to be properly executed and interpreted. Our results establish that reactions of NiI(0EiBC)- proceed by several competing pathways. These include reduction of alkyl halides to alkanes, coupling of alkyl halides to form new carboncarbon bonds, reaction with proton donors to afford H1,and the dehydrohalogenation of alkyl halides that have a @-hydrogen. The relative yields of these competing processes are senitive to the identities of the alkyl group and leaving group in the electrophilic reagent, the polarity of the solvent, and the level of residual proton donor (typically water) in the solvent. The evidence rules out free-radical processes as all but minor contributors to the mechanism and implies that carbon-hydrogen and carbon-carbon bond formation and dehydrohalogenation occur at nickel. Many of these processes formally require the availability of cis coordination sites on a metal. In this light, we suggest that the unusual structural preference of Ni(1) for two sets of different Ni-N distances could be functionally significant. T h e apparent failure of Ni'(0EiBC)- and NilF430M to react with thioethers, including methyl coenzyme M, is discussed in light of our results.

Experimental Section Owing to the extreme sensitivity of Nil(OEiBC)-, the following procedures were employed. All distillations were carried out under a nitrogen or argon atmosphere. Reductionsand subsequent reactions were performed under nitrogen inside a Vacuum/Atmospheres Co. drybox. (a) Gosden, C.; Healy, K. P.; Pletcher, D. J. Chem. Soc., Dalton Trans. 1978, 972. (b) Healy, K. P.; Pletcher, D. J. Urganomet. Chem. 1978, 161,109. (c) Gosden, C.; Pletcher, D. J . Organomet. Chem. 1980,186, Kerr, J. B.; Pletcher, D.; Rosss, R. J.Electroanal. 401. (d) Becker, J. Y.; Chem. Interfacial Electrochem. 1981, 117, 87. (e) Gosden, C.; Kerr, J. B.; Pletcher, D.; Rosas, R. J. Electroanal. Chem. Interfacial

Electrochem. 1981, 117, 101.

(a) Bakac, A.; Espenson, J. H. J. Am. Chem. SOC.1986,108,713. (b) Bakac,A.;Espenson,J. H. J. Am. Chem.Soc. 1986,108,719. (c) Ram, M.S.; Bakac, S.; Espenson, J. H. Inwg. Chem. 1986, 25, 3267. The authors of ref 1 1 suggest a different mechanism. See footnote 26 of ref 2.

Lahiri et al. The drybox was used exclusively for experiments with nickel(1) to prevent cross-contamination by volatile compounds released into the box atmosphere during unrelated experiments. All glassware was thoroughly cleaned, dried in a 120 "C oven overnight, and placed in the drybox antechamber while still hot. Spectrophotometric cells and serum bottles were treated with aqua regia prior to cleaning to prevent carry over of any traces of mercury from previous experiments. Materials. Ni(OEiBC)* and ammonium 2-(methy1thio)ethanesulfonate14(S-methyl coenzymeM) were prepared by literature methods. The tetraphenylphosphonium salt of the latter compound was obtained by metathesis of aqueous solutionsof the ammonium salt and tetraphenylphosphonium chloride. The precipitate was recrystallized from hot acetone and dried under vacuum for 3 days. Tributylammonium perchlorate was prepared by reaction of perchloric acid with tri-nbutylamine. It was recrystallized three times from hot anhydrousmethanol and thoroughly dried under vacuum at 100 "C. The resulting white, hygroscopic solid was stored in a dry environment. Reagents and solventsused in this study were HPLC or reagent grade. Alkyl halides and other reagents were examined for purity and were repurified by appropriate means if impurities were detected. Triphenylmethane was recrystallized from benzene and dried under vacuum at 60 "C. Tetrahydrofuran (THF) and diethyl ether were distilled from sodium benzophenone ketyl. N,N,N',N'-tetramethylethylenediamine (TMEDA) was purified by refluxing over KOH pellets followed by distillation. N-methylformamide (NMF) was dried over activated molecular sieves. 1-Methyl-2-pyrrolidinone (NMP) and propylene carbonate (anhydrous,under nitrogen) were used as received from Aldrich. Methanol was dried by refluxing over magnesium turnings that were activated with a crystal of 12 and was then distilled. Other solventswere refluxed over CaH2 and then distilled. All liquids and solvents were degassed by a minimum of three freeze-pumpthaw cycles. Five cycles were used for water and methanol. The water was treated previously by flushing with nitrogen gas overnight. General Procedure for Reductions and R e ~ c t i o ~Solutions . of NiI(0EiBC)- were prepared by reduction of solutions of NilI(OEiBC) with excess !% sodium mercury amalgam. The Nil*(OEiBC)solution contained a known weight of complex dissolved in a known volume of solvent that had been pretreated by stirring for 1 h with amalgam. The concentration of complex was typically 5 X M for experiments that were intended to monitor Nil(OEiBC)- by UV-vis spectroscopy and 2 X 10-3 M for experiments intended to measure the quantities of substratederived products. The solution and amalgam were stirred vigorously, and the progress of the reduction was followed spectroscopically by monitoring the decrease of the NiII(0EiBC) visible band near 592 nm. An aliquot of the solution was taken periodically and placed in a tightly stoppered cuvette. The cuvette was removed from the drybox and the spectrum was recorded. The time required for reduction and the stability ofthereducedspeciesdependedonthesolvent.Withexperience,itbccame possible to judge the completion of the reduction byeye. When reduction was complete, stirring was stopped and the finely divided amalgam was permitted to settle. Experiments intended to measure reaction products were typically run simultaneously in duplicate or triplicate. A calibrated to-deliver style pipet was used to dispense 5.00-mLaliquots of clear Nil(OEiBC)-solution intoseveral '10-mL" Wheaton serumbottles. (The totalvolumecontained in a capped "10-mL" Wheaton serum bottle is 15.0 mL.) Care was taken not to include any visible mercury. Consequently, a small volume of solution was not used. A Teflon-faced silicone septum (Wheaton no. 224173) and aluminum seal were placed on the mouth of each serum bottle, and the seal was crimped in place. A volume of liquid reagent that typically represented a 100-800-fold molar excess compared to nickel(1) was introduced by microliter syringe. Solid reagents, such as the alkyl p-toluenesulfonates, were dissolved in a minimal amount of pretreated solvent, and the resulting solution was added. The volume of added reagent did not significantlyalter thevolume of the reaction solution or head-space. The serum bottle was then shaken to assist equilibration of gases between solution and head-space and removed from the drybox. The products were promptly analyzed. The variation of results from duplicateserumbottles wassmall. Most experimentswererepeatedseveral times. Agreement between replicate experiments was good. Catalytic reactions were performed in the following manner. A known quantity of sodium amalgam, 5.00 mL of a solution of Ni(0EiBC) in pretreated solvent, and a small magnetic stir bar were sealed in a serum (14) Gunsalus, R. P.; Romesser, J. A,; Wolfe, R. S . Biochemistry 1978, 17, 2374.

Inorganic Chemistry, Vol. 31, No. 24, 1992 4993

F430 Model Chemistry bottle. The required quantity of the reagent of interest was added by syringe. The serum bottle was agitated to mix the contents and was immediately removed from the drybox. The sampling needle (see below) was then inserted through the septum of the serum bottle, and magnetic stirring was started. Head-space gases were sampled and analyzed immediately and at appropriate intervals thereafter. Product Analysis. Head-space gases and organic products were analyzed by gas chromatography or by GC/MS. A Hewlett-Packard 5890A gas chromatograph and 3396A integrator were used in these experiments. Separations were typically performed at temperatures between 35 to 40 OC. Compounds were identified by comparison of their retention times to thoses of pure compounds and mixtures of known composition. Low molecular weight hydrocarbons (C&) wereseparated on a 6-ft PoropakQ column. Methylcyclopentane,1-hexane,and 1-bromo5-hexene were separated on a 6-ft 10% OV-17 on chromosorb column. A flame ionization detector was used in both of theabovecases. Hydrogen was separated on a 6-ft type 5-A molecular sieve column and measured with a thermal conductivity detector. Argon was used as the carrier gas in hydrogen analyses to eliminate peak inversion or folding, which can result if helium is used as the carrier gas.ls The sensitivity of the detector to other components is greatly reduced by argon carrier gas. Methane and ethane could be separated and detected during hydrogen analyses, but the accuracy of the analyses for these gases was unsatisfactory. Thus, hydrogen and hydrocarbon products for a given reaction were determined separately using two sets of identically prepared solutions. The GC/MS analysisof deuteriated methanes and ethanes were conducted on a HewlettPackard Model 5990A that was fitted with a Poropak Q packed column. All columns were thoroughly conditioned prior to use and regenerated after each run by holding theoven temperatureat 200 OC until the solvent vapor peakcameoffthecolumn. Absorptionof solvent alters the properties of the column and increases the retention time of the hydrocarbon gases. This effect was responsible for our failure to detect ethane in the initial study. The precision of sampling head-space gases by gastight microliter syringe is too low to permit adequate determination of the mass balance of reactions at this scale. We used a gas sample valve system, whose design and operation are described in supplementarymaterial, to improve the precision of sampling head-space gases. The system enabled us to expand head-space gases into an evacuated, fixed-volume sample loop and then inject the sample onto the column. The reproducibility of injection of gas standards at constant pressure was &I%. The measured integrated peak area of each compound of interest in head-space gases was converted to a concentration by use of calibration curves that had been determined using analyzed gas mixtures purchased from Scott Specialty Gases, Inc. This method corrects for the unequal area response of flame ionization detectors for equal numbers of moles of different compounds.ISCThe number of moles of each compound that were present in the head-space was calculated from the concentration and the head-space, sample loop, and injector dead-space volumes. The error in the calculated moles of products is estimated at 5%, ignoring any loss of products through leakage or oxidationof Ni(1) during handling. The relative number of moles of different products within a single sample is known to a much higher precision. Detection limits are estimated to be roughly lo-" mol for hydrocarbon gases and 1O-Io mol for Hz. Extensive series of control experiments were performed to ensure that our results were not complicated by other factors. The septum material affected the recovery of products. Product recovery was highest when Teflon-faced silicone septa were used. Incident light had no effect on product yields provided that samples did not sit for an extended period of time (hours) before analysis. Direct reduction of substrates by sodium amalgam was much slower than the reaction with Ni'(OEiBC)-, with the exception of ethyl iodide. Carryover of amalgam or metallic mercury and their effect on product yields were both negligible. Direct sampling of reaction solutions suggested that slow equilibration between solution and head-space was not a problem. For the products analyzed here, the amount of product that remained in solution was small compared with the amount in the head-space. Finally, analysis of known quantitites of product gases produced in independent reactions showed that our anaytical and calibration methods gave reasonable results. Some control experiments are described in greater detail in the supplementary material. (15) (a) Cowper, C. J.; DeRose, A. J. The Analysis of Gases by Chromatography; Pergamon Press: Oxford, England, 1983. (b) Thompson, B. Fundamentals of Gas Analysis by Gas Chromatography; Varian Associates: Palo Alto, CA, 1977. (c) McNair, H.M.; Bonelli, E. J. Basic Gas Chromatography;Varian Associates: Palo Alto, CA, 1968.

RHult.9

Reduction of Ni(0EiBC) in Different Solvents. We investigated the reduction of Ni(0EiBC) by sodium amalgam in a range of solvents, as a preliminary to an examination of solvent effects on the reactivity of Nil(OEiBC)-. The anion is rapidly formed and is stable to further reduction for an extended time (>48 h) in 1,Zdimethoxyethane (DME), N,N-dimethylformamide (DMF), hexamethylphosphoramide(HMPA), and NMP. It is also stable in these solvents to reoxidation to Ni(I1) when separated from amalgam. In other solvents, the stability of the anion to reoxidation is somewhat less. In addition, Nil(OEiBC)-undergoes further reduction in these solvents to afford a reddish-brown colored, macrocycle reduced complex. The characterization of this complex will be reported elsewhere.I6 Reduction to Ni(1) is complete well before the reduction of the macrocycle begins in acetonitrile, THF, and 2-methyltetrahydrofuran (2-MeTHF). In tetrahydropyran (THP), the rates of the two reduction steps are sufficientlysimilar that quantitative reduction to Ni(1) cannot be achieved. Because the maximumconversion to Ni(1) achieved prior to formation of the macrocycle reduced complex was less than 90%, reported product yields in THP are artificially low. The reduction of the macrocycle was faster than the initial reduction to Ni(1) in NMF, 2,Sdimethyltetrahydrofuran (2,sMe*THF), and TMEDA. The macrocycle reduced complex was formed directly without buildup of a significant concentration of Ni(1). No reduction occurred in propylene carbonate. The last four solvents listed were not used in the reactivity studies that follow. In general, the rate of reduction and the stability of Ni'(0EiBC)- increase as the polarity of the solvent increases. A glaring exception to this is NMF, which is the most polar solvent by far (6 = 182 vs 36.7 for DMF). Unlike DMF, NMF has an amide proton, which can be removed by a base of moderate strength. This suggests that NiI(0EiBC)- could be involved in proton abstraction reactions. The reduction in HMPA requires further comment. Pretreatment of the solvent with sodium amalgam resulted in a yellow discoloration. When solid Ni(0EiBC) was added, dissolution of the solid and reduction occurred simultaneously in a matter of seconds. Reduction of Ni(0EiBC) in untreated HMPA took a few minutes. Alkali metals dissolve in HMPA to give solutions of the solvated e1ectr0n.l~ Addition of alkyl halides to these HMPA solutions affords Wurtz coupling products. Our observations suggest that sodium amalgam can give a small concentration of solvated electron in HMPA, too. Thus, amalgamdriven catalytic reactions were not studied in HMPA. Reaction of NiI(0EiBC)- with Hydrogen Donors. We reacted NiI(0EiBC)- with compounds that are either good donors of protons or of hydrogen atoms. The former category included water, methanol, 1-nitropropane,ammonium salts, and thiophenol. The later category included triphenylmethane and thiophenol. Reactionswere examined in several solvents. In all cases, addition of thedonor compound resulted in rapid, quantitative reoxidation of Ni(1) to Ni(II).l* We analyzed the head-space gases of several reactions for H2, Table I. The yield of Hz is essentially quantitative for the most acidic donor compounds and decrease as the pK* of the donor increases. Little or no hydrogen was observed in the reaction with triphenylmethane. Thus, Ni(1) can act as a base, albeit a weak base, and abstract H+ from acidic compounds. Nickel hydride complexes may be accessible and play a significant role in the reactions of Ni(0EiBC). (16) Lahiri, G. K.; Stolzenberg, A. M. Manuscript in preparation. (17) Fraenkel, G.; Ellis, S . H.; Dix, D. T. J . Am. Chem. SOC.1965,87, 1406. (18) One exception was the reaction with thiophenol in DMF. A small absorption band at 660 nm was observed after the reaction. This band cannot be reproduced by addition ofthiol, thiolate, or disulfide tosolutions of Ni"(0EiBC).

Lahiri et al.

4994 Inorganic Chemistry, Vol. 31, No. 24, I992 Table I. H2 Production from Ni'(0EiBC)THF HA CbHsSH H20 CH30H HN(C4H9)3CQ (C6HshCH

mol of H2/mol of Ni(1) 0.48 0.27 0.23 0.019

+ HAa DMF

mol of Hz/mol yieldb of Ni(1) 96 0.34 54 0.45 47 0.070 0.087 4 0 %

% yieldb 67 90 14 17 0

Typical conditions were 2 mM Ni(1). Liquid HA reagents were added neat, 600-3000 molar excess. Solid HA reagents were added as a solution in pretreated solvent, 50-70 molar excess. Assumes stoichiometry of 2 mol of Ni(1) consumed/mol of H2 produced. a

Table 11. Products of the Reactions of Ni'(0EiBC)- with CHjX, X = I or OTS".~

mol of product/mol of Ni(1) solvent

CH4

HMPA NMPf CH$N DMF THP 2-MeTHFg DME THF

0.10 0.11 0.098 0.074 0.052 0.22 0.12 0.42

HMPA' THF

0.22 0.23

C2H6

H2c

x = CI 0.12 0.077 0.055 0.033 0.022 0.0087 0.011 0.020 0.0037 0.0044 X = OTsh 0.082 0.00070

% yieldd C2:Cl ratio

44 37 31 21 12 46 28 85

1.2 0.70 0.56 0.45 0.17 0.050 0.17 0.0088

60 47

0.37 0.0030

a Typical conditions were 2 mM Ni(1). OTs = p-toluenesulfonate. H2 was not determined for solvents that have no listing. Yield of CH4 C2H6,assumes a stoichiometry of 2 mol of Ni(I)/mol of alkane. 800 mol of CH3I added/mol of Ni(1). f 1.O X IO-' mol of C2H4 also detected. 8 4.9 X mol of C2H4 also detected. 4 mol of CH30Ts added/mol of Ni(1). i 7.1 X lo4 mol of C2H4 also detected. c

+

Reaction of Ni*(OEiBC)- with CH3X, X = I or OTs. NiI(OEiBC)-is oxidized to NiII(0EiBC) upon mixing with CHJ or CH30Ts. Although the relative rate for reaction with CH30Ts is smaller,2 both reactions have kinetics that are much too fast to be studied by conventional techniques. For both substrates, the principal products are methane and ethane. Lesser amounts of hydrogen are observed. Trace quantities of ethylene (roughly 5X to 1 X mol per mole of Ni(1)) are observed in some solvents. (i) Effect of the [CHJX]/[Ni(I)] Ratio. The effect on product yields of the [CH3X]/[Ni(I)] ratio at fixed [Ni(I)] (2 mM) was examined in THF and HMPA. Relatively minor changes occurred. For CH31 in THF, an increase in the reactant ratio from 4 to 950 resulted in a 30% increase in yield of methane, a 3-fold increase in yield of ethane, and no change in the yield of ethylene. In this solvent, an increase in the reactant ratio favors the coupling reaction. No significant changes in product yields were observed in HMPA when the [CH31]/[Ni(I)] ratio was varied over the same range. For CH30Ts in both solvents, the yield of all products fell by roughly 40% when the ratio was varied from 3 to 650. If solid CH30Ts contains some water, the drop in yield could reflect an increase in the level of water that was introduced into the reaction as a result of increasing the amount of CH3OTs added. On the basis of these results, we decided to analyze products yields in reactions with large excesses of CH31 but small excesses of CH3OTs. (ii) Solvent Effects on Product Distributions. Product distributions from the reactions of Ni'(0EiBC)- with CH31 and CHjOTs in a rangeof solvents are presented in Table 11. Because percentage yields of products depend on the assumed reaction mechanism, we also report the number of moles of product produced per mole of Ni(1).

The mass balance of the reaction is essentially complete for CH31 in THF. Given the potential for partial reoxidation of Ni(1) and loss of product gases through the once-punctured septum, an 85%combined yield of methane plus ethane (assuming 2:l stoichiometry for both products) should be considered close to quantitative. Yields are lower in other solvents. The analysis of H2 produced during the reaction in DMF establishes that the mass balance is not completed by production of large quantities of H2. We did not detect any other volatile product in the lower yield reactions. The missing products could be solvent derived species that do not distribute into the head-space gases. Alternatively, partial reoxidation of Ni(1) by adventitious oxidants could account for the lowered yields. The ratio of ethane to methane produced reflects the relative rates of coupling reactions and reduction reactions. The ratio shows a large solvent dependence, changing by a factor of 136 between the solvents HMPA and THF. The relative importance of coupling reactions appears to correlate directly with the polarity of the solvent. Previously. we remrted that the yield of methane was smaller with CH3OTs than with CH31.2 T k s observation is true in THF but is not general in all solvents. The reaction of NiI(0EiBC)with CH3OTs is somewhat slower than the reaction with CH31, and the mass balance is less complete. The ratio of ethane to methane formed from CH3OTs is roughly 0.25 that observed for CH31. Thus, theidentityof thexgroupof theelectrophileaffects the distribution of products. (iii) Low-Temperature Reaction of NiI(0EiBC)- with CHJI. Previousattempts to detect the proposed alkyl-nickel intermediate CHjNi(0EiBC) at room temperature were unsuccessful. We examined the reaction at a lower temperature to see if the intermediate could be stabilized. Solutions of Ni'(0EiBC)- in either THF or DMF were transferred to a quartz optical dewar. The solutions were cooled to -41 OC by filling the dewar with an acetonitrile/liquid-N2slush. The UV-vis spectra of Ni*(OEiBC)did not change upon cooling. Addition of cold CHJ resulted in immediate, quantitative reoxidation to NP(0EiBC) as judged by the UV-vis spectrum. In separate experiments, analysis of the head-spacegas over reactions conductedat -41 OC established that there was no detectable delay between CH3I addition and methane format ion. (iv) IsotopicLabelingExperiments. We examined the reactions of NiI(0EiBC)- with deuterium and IT-labeled CH31in normal and deuteriated solvents. The distribution of isotopes in the products was determined by use of an HP 5990A GC/MS. Background levels of 0 2 , N2, and H20 were relatively high in the instrument available to us. Masses 32, 28, and 18 were totally obscured. The fragmentation patterns and ionization cross section of individual deuteriated methanes and ethanes varies with the compound and with the conditions in the ion source of the mass ~pectrometer.'~ Thus, we did not attempt to deconvolute the observed mass spectra into quantitative distributions of individual deuteriated compounds. Nonetheless, our qualitative estimates are sufficiently good to draw several important conclusions. Reaction of Ni'(0EiBC)- with CDJ in THF afforded methane that had a mass spectrum dominated by a peak at mass 19, which corresponds to CD3H. This suggests that excess CD31is not the source of the fourth hydrogen atom in methane. The distribution of masses for ethane were reasonable for the expected fragmentation of C2D6.19.20Thus, ethane is a product of methyl iodide and does not result from cleavage of an ethyl substituent from the isobacteriochlorin. Ethylene was not detected above the instrument background. (19) (a) Miller, T. M.; McCarthy, T. J.; Whitesides, G. M. J . Am. Chem. SOC.1988, 110, 3156. (b) Miller, T. M.; Whitesides, G. M. J . Am. Chem. SOC.1988,110, 3164. (c) Lee,T. R.; Whitesides, G. M. J . Am. Chem. SOC.1991, 113, 2568. (20) Amenomiya, Y.; Pottie, R. F. Can. J . Chem. 1968.46, 1735.

Inorganic Chemistry, Vol. 31, No. 24, 1992 4995

F430 Model Chemistry An experiment with CD31was conducted in THF-dg that had been pretreated with sodium amalgam. The mass spectrum of methane had comparable abundances of masses 19 and 20. The product contained both CD3H and CDd. Similarly, reaction of NY(0EiBC)- with CHJ in pretreated THF-dg afforded comparable quantities of CH4 and CHjD. Both experiments were repeated using glassware that had been washed with D2O and then thoroughly dried. The results were essentially unchanged. Thus, one source of the fourth hydrogen in methane is the solvent, but a pool of protons remains available even in deuteriated solvent. Two candidates for the pool of protons are residual water and possibly the isobacteriochlorin ligand, although the latter seems unlikely given the apparent quantitative recovery of Ni(0EiBC). An experiment conducted with l3CH31in THF gave a peak at mass 17 for methane and a large mass 31 peak for ethane (mass 32 was obscured by background 02). Ni(0EiBC) was recovered from the reaction by stripping off all volatiles. The resulting solid was dissolved in CDC13 and examined by NMR. No new or enhanced 13Cpeaks were observed. The spectrum was identical to that for natural-abundance Ni(0EiBC). Thus, a methyl iodide derived carbon is not incorporated in any modified OEiBC compound. (v) CatalyticReactions. Solutions of NiII(0EiBC) in pretreated CH3CNor THF were reacted with approximately 100-200 equiv of sodium amalgam (Le. mol Na/mol Ni) and excess methyl iodide. The color of the solution remained that of Ni"(OEiBC) throughout the reaction, which suggests that reduction to Ni(1) is the slow step in the catalytic cycle. Head-space gases were sampled periodically to follow the progress of the reaction. The mole ratios of products gases appeared to stay roughly constant during the reaction while the total number of moles of each product increased with time. The amount of methane produced in experiments conducted in CH3CN suggests that the reducing equivalents in the amalgam are converted to products with relatively little loss to other processes. The conversion efficiency in THF is substantially less or else thereduction stops prematurely. Product ratios in catalytic and batch reactions are not identical. The ethane to methane ratio drops from 0.56 in batch reactions to 0.030 in catalytic reactions in CH3CN. In contrast, the ratio increases from 0.0088 in batch to 0.54 in catalytic reactions in THF. The amount of ethylene produced in catalytic reactions is small. The ratio of ethylene to methane is about 2 X in THF but only about 7 X lo4 in CH3CN. These observations imply that either the course of the catalytic and batch reactions are different or that changes in relative concentrations of the reactive species changes the partitioning between competing reactions. Reaction of Nil(0EiBC)- with CzH&, X = I or OTs. Product distributions from the reactions of NiI(0EiBC)- with C2HJ and C2HsOTs in several solvents are reported in Table 111. The principal products are ethane, ethylene, and butane. Small amounts of butenes are formed in some solvents. The table lists moles of product produced per mole of Ni(1) and the ratios of ethylene to ethane and of ethylene to butene. Butane was not determined in the experiments in CH3CN and in 2-MeTHF because its peak was obscured by the solvent peak. Overall, product yields are significantly lower for C2HsIthan for CH31. One exception to the low yields occurs in HMPA. The total yield of products in this solvent is greater than expected. In particular, much more ethylene is formed in HMPA than in other solvents. The high polarity of the HMPAcould promote a solventassisted, non-Ni(1)-dependent dehydrohalogenation pathway. The product ratios exhibit significant solvent dependence. The ethylene to ethane ratio in DMF is approximately 1.O,the value expected for self-reaction of ethyl radicals. However, in other solvents the ratio ranges from 0.025 to 3.6. The ratio of ethylene to butane is generally greater than 1.0 and in some solvents is very large. In general, both product ratios are largest in polar

Table 111. Products of the Reactions of NiI(0EiBC)- with C~HSX, X = I or OTsaqb

HMPA NMP CH3CN DMFf THP 2-MeTHF DMEg THF".'

x= 0.15 0.059 0.028 0.046 0.027 0.11 0.039 0.070

0.53 0.074 0.036 0.047 0.0024 0.0027 0.034 0.013

IC

0.010 0.0027 d 0.0068 0 d 0.0025 0.0038

3.6 1.3 1.3 1 .o 0.089 0.025 0.86 0.19

52 27 e 6.9

large e 14 3.5

X = OTs 0.015 0.00062 0.0015 0.042 DMFj-' 0.42 THFm 0.027 0.00067 0.00062 0.025 1 .o a Typical conditions were 2 mM Ni(1). OTs = p-toluenesulfonate. c 800 mol of C2H51added/mol of Ni(1). Obscured by solvent peaks. e Could not be determined. f 0.072mol of H2 observed. 8 1.4X 10-3 mol of C4H8 observed. * 5.5 X mol of C4H8 observed. Checked for H2; mol none observed. j 180mol of EtOTs added/mol of Ni(1). 2.4 X of C4H8 observed. 0.033mol of Hz observed. 68 mol of EtOTs added/ mol of Ni(1).

solvents. Dehydrohalogenation and @-hydride elimination2' involve polar transition states and are favored by an increase in the polarity of the solvent. Reactions of NiI(0EiBC)- with C2H5OTs give significantly less ethylene than reaction with CZH51.Coupling to form butane is reduced slightly, too. Again, the identity of the X group of the electrophile affects the distribution of products. Reactions of NiI(0EiBC)- with Hydrogen Donors and CHJ. The deuterium labeling experiments described above establish that a pool of hydrogen distinct from that in the solvent takes part in the reaction. The most probable candidate for the pool is the residual water in the "dry" solvent. The concentration of residual water could affect the course of the reaction and the distribution of products. Normally, to address this issue, one would add a known concentration of water or hydrogen donor and determine how it affected the distribution of products. Such experiments are not straightforward here. Addition of a hydrogen donor prior to the reduction of Ni(0EiBC) either prevents the reduction, leads to consumption of the donor, or results in formation of the macrocycle-reduced complex. On the other hand, addition of hydrogen donors to NiI(0EiBC)- results in reoxidation to Ni"(0EiBC) and formation of H2. One possible way to perform the experiment is to simultaneously add CH3I and the hydrogen donor in the hope that reduction of the alkyl halide and formation of HZoccur in competition. Equimolar mixtures of CHJ and a proton donor were added in a 300:l molar excess (per reactant) to NiI(0EiBC)- in THF or DMF. (CH31and H2O were added as a two-phase mixture.) The product distributions are reported in Table IV. The two reactions compete with each other. Comparisons of Table IV with Tables I and I1 show that the yields of both reactions are significantly lower when run simultaneously than when run independently of each other. Addition of a proton donor decreases the yield of methane in THF by a factor of 4 or more. The presence of CH31reduces the yield of H2 by only about one-third for C6H5SHbut by more than a factor of 9 for CH30H and H20. A far more significant change is that the added proton donors completely suppress the coupling reaction that affords ethane from CH31. Proton donors do not have that large an effect on the actual yield of methane in DMF. "Dry" DMF is typically 1-2 mM in water.22 The amount of proton donor that we added (21) (a) Doherty, N. M.; Bercaw, J. E. J . Am. Chem. SOC.1985,107,2670. (b) Burger, B. J.;Thompson, M. E.; Cotter, W. D.; Bercaw, J. E.J. Am. Chem. SOC.1990,112, 1566. (22) La Perriere, D. M.; Carroll, W. F., Jr.; Willett, B. C.; Torp, E. C.; Peters, D. G. J. Am. Chem. SOC.1979, 101, 7561.

4996 Inorganic Chemistry, Vol. 31, No. 24, 1992

Lahiri et al.

Table IV. Effect of HA on the Products of Reactions of Ni'(0EiBC)- with C H 3 P b

Ni(1) to Ni(II), but only small quantities of methane mol/ mole Ni) were produced. We prepared the tetraphenylphosphonium salt of methyl CoM by metathesis in order to prevent oxidation of Ni(1) by the acidic ammonium cation. Reaction of tetraphenylphosphoniummethyl CoM with NiI(OEiBC)-in DMF resulted in immediateoxidation of Ni(I1). No methane was detected. The catalytic reduction of tetraphenylphosphonium methyl CoM was also attempted. Any impurity in the methyl CoM salt that oxidizes Ni(1) more rapidly than Ni(1) can react with methyl CoM should eventually be consumed in the catalytic reduction. The reaction of Ni(1) with methyl CoM would then be observable. The reduction of Ni(0EiBC) in the presence of excess tetraphenylphosphonium methyl CoM proceeded eventually to Ni(1). No methane was detected during or after the reduction. Reaction of a mixture of tetraphenylphosphonium methyl CoM and C ~ H S S Hwith Ni'(0EiBC)-in DMF led to reoxidation to Ni(I1) and production of Hz, but no methane was formed.

HA

mol of product/mol of Ni(1) CHI C2H6 H2

C6HsSH H20 CHjOH

0.10 0.10 0.071

C6HsSH H20 CH3OH

0.096 0.065 0.054

THF 0 0.3 1 0 0 0 0.017 DMF 0.052 0.025 0.026 0.26 0.012 0.018

% yieldc

C2:C) ratio

20 20 14

0.0 0.0 0.0

24 18 14

0.26 0.42 0.33

*

Typical conditions were 2 mM Ni(1). 300 mol each of CH31 and HAadded/molof Ni(1). Yieldof CH4 + C2H6,assumingastoichiometry of 2 mol Ni(I)/mol of alkane. Table V. Products of the Reactions of Ni'(0EiBC)- with 1-Bromo-5-hexenea

relative yields solvent CH3CNC DMF THP DME 2-MeTHF DMF DMF DMF DMF

HAb

1-hexene

C6HsSH CHjOH D20d H20d

60 74 81 90 92 74 82 81 93

methylcyclopentane 40 26 19 10 8 26 18 19 7

Typical conditions were 2 mM Ni(1) plus 200 molar excess of 1-bromo5-hexene. 220 molar excess of HA added simultaneously with bromohexene. Reference 2. Total yield decreases by about 45% relative to yield in DMF without addition of HA. (I

may not change the concentration of available proton sufficiently to cause a significant decrease in the methane yield. The ratio of ethane to methane is decreased though, especially by ChHsSH and CH30H. By way of contrast, CH31 causes a significant reduction in the yield of H2 from the proton donors in DMF. Reactions of NiI(0EiBC)- with 1-Bromo-5-hexene. We reported in our initial study that the reaction of Nil(0EiBC)- with excess 1-bromo-5-hexenein CH&N afforded a 60:40 mixture of 1-hexene and methylcyclopentane.2 In light of the sensitivity of product distributions to the identity of the solvent and the presenceof hydrogendonors, we decided to reexaminethe reaction. The relative yields of 1-hexene and methylcyclopentane in several dry solvents are reported in Table V. The table also includes data for experiments in which hydrogen donors and 1-bromo-5-hexene were added simultaneously to Ni'(0EiBC)in DMF. 1-Hexene was the major product in all solvents examined. The solvents that afford greater amounts of methylcyclopentane are those that afford more ethane from CH31. The product distribution of the reaction in DMF is unaffected by the presence of C6HsSH. In contrast, the presence of CH3OH and H2O affect the product distribution. The relative yield of 1-hexene climbs from 74% to 93% when H2O is added. Interestingly, D20 causes a much smaller increase in the amount of 1-hexene, which may reflect a kinetic isotope effect. The total integrated peak area of products decreases when water is added. Reactions of NP(0EiBC)- with Methyl CoM. The natural substrate of methyl coenzyme-M reductase is the 2-(methylthi0)ethanesulfonate anion, methyl CoM. The anion is usually isolated as the ammonium salt, which has limited solubility in nonaqueous solvents. Reaction of ammonium methyl CoM with NiI(0EiBC)- in DMF resulted in the immediate oxidation of (23) (a) Ellermann, J.; Kobelt, A,; Pfaltz, A.; Thauer, R. K. FEBS Lett. 1987,220, 358. (b) Bobik, T. A,; Olson, K. D.; Noll, K. M.; Wolfe, R.

S.Biochem. Biophys. Res. Commun. 1987, 149, 455.

Discussion A satisfactory mechanism for the reactions of NiI(0EiBC)with electrophilic reagents must account for the following observations: (1) Reactions with alkyl halides and tosylates follow a mixed second-order rate law. (2) The stoichiometry of the reaction with CH31 is 2:l Ni(1):CHJ. (3) The reactivity order follows the trends RI > RBr > ROTS > RCl and CH3Br >> n-C4H9Br> s-CqH9Br > t-C4H9Br. (4) Ni(0EiBC) mediates electrocatalytic reduction of alkyl halides and tosylates at the potential of the Ni(II)/Ni(I) couple. ( 5 ) Ni'(0EiBC)- reacts with proton (H+) donors to afford HZ with an apparent stoichiometry of 2:l Ni(I):Hz. (6) Alkyl halides and tosylates afford both reduction (RX RH) and coupling (RX R-R) products. (7) Alkyl halides and tosylates that have j3-hydrogens also undergo elimination to afford alkenes (RX R-H). (8) The relative yields of products shows a strong solvent and leaving group dependence. Polar solvents favor both coupling and alkene formation. (9) The fourth hydrogen atom in methane is derived from the solvent and from an unidentified proton pool (presumed to be residual water) in the solvent. (10) The recovery of NiII(0EiBC) is quantitative within the limits of detection, even after multiple turnovers. (1 1) Simultaneous addition of alkyl halide and proton donor increases the yield of reduction products relative to coupling products. In the case of 1-bromo-5-hexene, simultaneous addition of proton donor increases the yield of the reduction product 1-hexenerelative to the rearrangment product methylcyclopentane. (12) Trace levels of ethylene and butene are produced from CH3X and C ~ H S Xrespectively. , The mechanism outlined in eqs 2-14 accommodates all of the observations listed above, with the exception of the last one. We will defer a discussion of this observation until later in this paper. The symbol e- designates a reducing agent, which can be either an electrode poised at a potential 5-13 V vs SCE or a chemical reducing agent such as sodium In batch reactions, eqs 4 and 12 do not occur and eq 2 is temporally and spatially separated from eqs 3, 5-1 1, 13, and 14. In catalytic reactions, eq 2 occurs simultaneously to and in the same vessel as eqs 3-14. Theoxidation states indicated for nickel are the formal oxidation states that result if one follows the convention of counting an alkyl or H group as a closed-shell anionic ligand.24 OEiBC is a dianionic ligand. Studies of electrocatalytic reduction of alkyl halides by Ni(0EiBC) establish that eq 3 is second order and rate limiting.2 Given the above observations, a 17-e- paramagnetic alkylnickel complex R-NiII'(OEiBC) is the most reasonable product of eq 3. An intermediate of this or of some other type must exist to

-

-

-

(24) Collman, J. P.; Hegedus, L.S.;Norton, J. R.; Finke, R. G.Principles and Applications of Organotransition Metal Chemistry; University Science: Mill Valley, CA, 1987.

F430 Model Chemistry Ni"(0EiBC) Ni'(0EiBC)-

+ R-X

R-Ni"'(0EiBC) R-Ni"'(0EiBC)

+ + e-

Inorganic Chemistry, Vol. 31, No. 24, 1992 4997

Ni'(0EiBC)-

R-Ni"'(0EiBC)

e-

+ Ni'(0EiBC)-

R-Ni"(0EiBC)R-Ni"(0EiBC)R-Ni"(0EiBC)-

+ HA + R-X

R-Ni"(0EiBC)R-Ni"'(0EiBC) Ni'(0EiBC)-

+ R-X

-

+ X-

R-Ni"(0EiBC)-

-

+ Ni"(0EiBC)

+ Ni"(0EiBC) + AR-R + Ni"(0EiBC) + X-

R-H

(6)

(7)

H-Ni"'(0EiBC)

+ R-H

(9)

+ R-H + X-

+ HA +H-Ni"'(0EiBC) + AH-Ni"'(0EiBC) + eH-Ni"(0EiBC)-

-

+ Ni'(0EiBC)-

-+

H,

(10) (1 1) (12)

2Ni(OEiBC')+

+ HA

-

-

R-H Ni(0EC)

Ni(OEiBC')+

+ Ni"'(OEiBC)+ + A-

+ Ni(0EiBC) + 2H+

+ Ni'(0EiBC)-

Ni(OEiBC')+

+ e-

-

-

(15) (16)

2Ni(OEiBC) (17)

Ni(0EiBC)

(18)

preferential loss of an electron from the OEiBC *-system rather than from Ni(II).29 NilI1(OEiBC)+,the product of eq 15, is an excited state of the Ni(OEiBC*)+cation radical. If formed, it should relax to the cation radical, which is unstable in DMF and quickly disproportionates, eq 16.29 It is difficult to assert that both eqs 17 and 18 are sufficiently rapid in all circumstances to the prevent formation of any Ni(0EC) in DMF solution. The observed quantitative recovery of Ni(0EiBC) requires this to be true if eq 15 plays a significant role in the mechanism. Homolysis of R-Ni(OEiBC), eq 19, cannot be ruled out. However, several observations are inconsistent with a major role R-Ni"'(OEiBC)

-

Ni"(0EiBC)

+ R'

(19)

Ni"(0EiBC) (1 3)

H-Ni"(0EiBC)-

-

(5)

(8)

-

+ HA

R-Ni"'(0EiBC)

+ R-H

Ni'(0EiBC)-

H-Ni"(0EiBC)-

(4)

H-Ni"(0EiBC)-

H-Ni"'(0EiBC)

H-Ni"'(0EiBC)

(3)

potentials is also reasonablebecause the 17-e-alkylnickelcomplex has a hole available in a low-lying orbital. We intentionally omit protonolysis of R-NilI1(OEiBC), eq 15, from the mechanism. Oxidation of Ni(0EiBC) results in

(2)

+ Ni"(0EiBC) + A-

-

(14)

2C2H,' C,H,, explain the 2: 1 stoichiometry of the overall reaction. Our inability to find direct spectroscopic evidence for R-Nittl(OEiBC) at R' Ni'(0EiBC)R-Ni"(0EiBC)(23) temperatures as low as -41 OC does not rule out the complex as the intermediate. UV-vis spectroscopy may not be sensitive to in the mechanism for homolysis followed by free-radical reactions the formation of a C-Ni bond. Furthermore, the intermediate@) of the resulting alkyl radicals, eq 20. A mechanism consisting must be very reactive because the reaction is essentially complete of eqs 2, 3, 19, and 20 has a stoichiometry of 1:l Ni(I):CH31 upon mixing at 4 1 OC. Stable alkylnickel tetrapyrrole comrather than the observed 2:l stoichiometry. If the products of plexes are not known. However, for other metals, reaction the reactionsof NiI(OEiBC)-with alkyl halides are formed solely of a low-valent metallotetrapyrrolecomplex with an alkyl halide (or predominantly) via free radicals, the product distributions in affords stable alkyl-metallotetrapyrole complexe~.~~ Of greater these reactions must be identical (or similar) to those determined relevance are two reports of alkylnickel complexes of ligands for free radicals generated by independent means. The ratio of with an Nd donor set. [CH3Ni*I(tmc)][CF3SO3] is sufficiently the rate constants for disproportionation and comproportionation stable that it can be isolated.26 A paramagneticmethylnickel(I1) reactions of ethyl radicals, eqs 21 and 22, respectively, has been F430M complex was prepared at -78 OC and characterized by reported for ethyl radicals prepared by several independent NMRS2' The complex decomposes slowly at -20 OC. methods in different sol~ents.~O The ratio, which equals the ratio R-Ni(0EiBC) can be consumed in several reactions. If R has of ethylene to butane even if other pathways that produce ethane a @-hydrogen,it could undergo &hydride elimination, eq 9. exist, is not strongly affected by the reaction medium. It varies R-NilI1(OEiBC) can also be reduced, eq 4 or 5. The standard from 0.13 in vacuum to 0.35 in water. This is inconsistent with reduction potentials of alkylnickel complexes are not known. our results for the reduction of ethyl iodide. Our observed ratios However, the observation that electrccatalysis occurs at the of ethylene to butane are significantly larger and show a strong potential of the NiII(OEiBC)/Ni'(OEiBC)- couple requires that solvent dependence. It is hard to rationalize how the addition of the standard potential of the R-Nittl(OEiBC)/R-NiI1(OEiBC)proton (H+) donors would increase the yield of methane but couple be similar to or positive of the potential of the former decrease the yield of ethane from methyl free radicals. Finally, couple. If not, a separate reduction wave would be observed for the predominance of nonrearranged products in the reductions the alkylnickel couple at a more negative potential and the onset of l-bromo-5-hexene requires that any free radicals formed be of catalytic reduction in the presence of excess alkyl halide would rapidly captured, as in eq 23. Radical capture reactions of this occur near the potential of this second wave.28 The ordering of type were proposed in thereactionsof NiI(tmc)+with alkyl halides and were estimated to have rate constants approaching 108 M-1 (25) (a) Brothers, P. J.; Collman, J. P. Acc. Chem. Res. 1986, 19, 209. (b)

+

~

~

-

~~

Guilard, R.; Lecomte, C.; Kadish, K. M. Srrucr. Bonding 1987,64,205. (c) Guilard, R.; Kadish, K. M. Chem. Rev. 1988, 88, 1121. (26) D'Aniello, M. J., Jr.; Barefield, E. K. J . Am. Chem.Soc. 1976,98,1610. (27) Lin, S.-K.; Jaun, E. Helu. Chim. Acra 1991, 74, 1725. (28) Lexa, D.; Saveant, J.-M.; Wang, D. L. Organomerallics 1986,5, 1428.

(29) Stolzenberg, A. M.; Stershic, M. T. Inorg. Chem. 1988, 27, 1614. (30) (a) Sheldon, R. A.; Kochi, J. K. J . Am. Chem. Soc. 1970,92,4395. (b)

Gibian, M. J.; Corley, R. C. Chem. Reu. 1973,73,441. (c) Bakac, A.; Espenson, J. H. J . Phys. Chem. 1986, 90, 325.

4998 Inorganic Chemistry, Vol. 31, No. 24, 1992 s-I.12a Note that the sum of eqs 19 and 23 represents the same net change as eq 5. The proposed product-forming reactions include eq 6, which affords the reduction product, eq 7, which affords the coupling product, and for R groups that have a @-hydrogen,eq 8, which affords the HX elimination product. These are respectively protonation, alkylation, and internal proton-transfer reactions of R-NiIl(OEiBC)-. The product distribution reflects the relative rates at which the three reactions consume this common intermediate. Because eqs 6 and 7 follow mixed second-order kinetics and eq 8 follows first-order kinetics, the product distribution should be sensitiveto the identitiesand concentrations of RX and HA and to the identity of the solvent. This is consistent with our observations. The rate of eq 7 and the relative amount of coupled product should increase with the reactivity of the electrophile RX. As expected, the relative yield of ethane is higher for the more reactive CHJ than for CH3OTs. Similarly, the relative yield of butane is greater for C2HJ than for C2HsOTs. For reactions in "dry" solvents, the concentration of HA is set by the level of residual water and the availability of protons from the solvent. Intentional addition of an acidic compound will increase this concentration and should favor the rate of eq 6 over thatofeq7. Simultaneousadditionof CHJandacidiccompounds does increase the yield of methane (reduction) relative to ethane (coupling), Table IV. R-NiIl(OEiBC)- functions as a base and as a nucleophile, respectively, in eqs 6 and 7. The complex is expected to be both a strong base and nucleophile. R-Nill(OEiBC)- is isoelectronic with CH3-Ni1I( F430M) and CH3-Nil1(tmc)+. The three complexes differ in overall charge because of the charges of their macrocyclic ligands are dianionic, monoanionic, and neutral ligands, respectively. The difference in charge has a substantial affect on the reactivity of the tmc and F430M complexes. CH3NiII(tmc)+ hydrolyzes in alkaline aqueous solution (9 < pH < 13)onatimescaleof 1-15minat 25 oC.12a,31 CH3-Ni11(F430M) decays slowly at -20 OC in nonaqueous solution, presumably by irreversible formation of methane after proton transfer from an amide or lactam group of F430M.27 Addition of acid results in immediatehydrolysis of CH3-NiI1(F430M) and CH3-Ni11(tmc)+. Given its negative charge, R-NiII(OEiBC)- should be even more reactive. To this point in the discussion, the reactions that produce alkenes, eqs 8 and 9, involve intermediates that are common to reactions that afford other products. Alkene formation is also possible by an independent mechanism. Ni'(0EiBC)- could function as a base in an E2 elimination of HX from RX, eq 10. Reduction of protons to H2 by Ni(1) tetraazamacrocyclic complexes has been observed.32 The mechanism is thought to involve protonation of the Ni(1) complex to afford a hydride Ni(II1) complex.32-33 We invoke an analogous step, eq 11, in the mechanism of H2 production by Nil(OEiBC)-. Reduction of the resulting H-NilI1(OEiBC), eq 12 or 13, followed by a second protonation, eq 14, affords Hz. The thermodynamic feasibility of eqs 12 and 13 can be supported by arguments similar to those used in the case of eqs 4 and 5. The equilibrium in eq 11 appears to favor NiI(0EiBC)- unless HA is a strong acid or present in large excess. Consequently, H-NiIf1(OEiBC)- that is produced in eqs 8-10 can dissociated and regenerate NiI(OEiBC)-. In general, Ni(1) complexes do not appear to be particularly strong bases toward protons. Nil(tmc)+ is stable for hours in alkaline (pH = 11) aqueous solution.12 Electrocatalyticreduction of C02 competes effectively with H2 production when nickel tetraazamacrocyclic complexes are reduced under a C02 atmosphere in weakly acidic (pH = 4)

(31) Ram, M. S.; Espenson, J. H.; Bakac, A. Inorg. Chem. 1986, 25, 4 1 15.

Lahiri et al. aqueous solution.32 This discrimination toward substrates is remarkable given the weak coordinating ability of C02, the large concentration of proton in solution, and more favorable thermodynamics for H2 production over C02 reduction. Relationship of Reaction Mechanism and the Structural Preferences of Ni(1). The discussion above does not address the detailed mechanisms of eq 6,7, or 14. Mechanisms that require initial cleavage of the C-Ni bond can be neglected. Arguments presented against product formation via eqs 19 and 20 apply equally well to homolysis to R-Nill(OEiBC)-. Heterolysis to yield Ni(0EiBC) and R- is unreasonable because R- has much too high a free energy to be formed as an intermediate in solution.31 Two potential types of mechanisms for eqs 6 and 7 differ with respect to the polarization of the C-Ni bond. The a-carbon of the alkyl group has significant carbanionic character in mechanisms of the first type. Equation 6 proceeds by formation of a C-H bond between the a-carbon and the proton and simultaneous cleavage of the C-Ni bond. Similarly, eq 7 involves a nucleophilic attack by the a-carbon on R-X and proceeds with compensating bond making (C-C) and bond breaking (C-Ni and C-X) processes. In mechanisms of the second type, the nickel atom is the nucleophilicsite. Equation 6 involves protonation of the nickel atomof R-Ni"(OEiBC)-, eq 24, followed by reductive elimination, eq 25. Equation 7 also proceeds in two steps: alkylation of R-Ni"(OEiBC)- to afford a dialkylnickelcomplex, eq 26, followed by reductive elimination to yield the coupled alkane product, eq 1 . )

L I .

R-Ni"(OEiBC)-

+ HA

(R)(H)Ni"(OEiBC)

(R)(H)Ni"(OEiBC) R-Ni"(0EiBC)-

+ R'-X

+

-

R-H

+ Ni"(0EiBC)

(R)(R')Ni'"(OEiBC)

(R)(R')Nilv(0EiBC)

R-R'

+ A-

(24) (25)

+ X- (26)

+ Ni"(0EiBC)

(27)

Precedents exist for organonickel complexes reacting by both types of mechanisms. The key intermediates in the nickelcatalyzed coupling of aryl halides and cross coupling of alkyl and aryl halides are diarylnickel(II1) and arylalkylnickel(I1) complexes, re~pectively.~~ On the other hand, hydrolysis of R-Ni(tmc)+ complexes was suggested to occur by direct reaction of the a-carbon with the proton donora3I Organonickel complexes NC(CH&CH2Ni(tmc)+ and CH2=CH(CH2)sCHzNi(tmc)+ undergo cyclization of the alkyl group in concert with hydrolysis to afford cyclopentanone and methylcyclopentane, re~pective1y.l"~~~ The cyclization is presumed to involve a nickel-bound carbanion, in analogy to the cyclizationof hexenyl alkali metal c0mpounds.3~ Two mechanisms can be proposed for the reaction of H-Ni"(OEiBC)- in eq 14 that are analogous to the possibilities discussed above for R-NiII(OEiBC)-. They differ with respect to the polarization of the Ni-H bond. Protonation on nickel, eq 28, affords a classical dihydride complex, which then undergoes reductive elimination to afford H2, eq 29. Alternatively, pro(32) (a) Fisher, B.; Eisenberg, R. J. Am. Chem. SOC.1980, 102, 7361. (b) J. Am. Chem. Soc. Beley, M.; Collin, J.-P.; Ruppert, R.; Sauvage, J.-P. 1986, 108, 7461. (33) Tait, A. M.; Hoffman, M.Z.; Hayon, E. Inorg. Chem. 1976,15,934. (34) (a) Morrell, D. G.; Kochi, J. K. J. Am. Chem. SOC.1975,97,7262. (b) Tsou, T. T.; Kochi, J. K. J. Am. Chem. SOC.1979, 101, 7548. (c)

Yamamoto, T.; Kohara, T.; Osakada, K.; Yamamoto, A. Bull. Chem. SOC.Jpn. 1983.56.2147. (d) Amatore, C.; Jutand, A. Organometallics 1988, 7, 2203. (35) Bakac, A.; Espenson, J. H. J . Am. Chem. SOC.1986, 108, 5353. (36) (a) Garst, J. F.; H i m , J. B., Jr. J. Am. Chem. SOC.1984, 106, 6443. (b) Bailey, W.F.;Patricia, J. J.; DelGobbo,V.C.;Jarret, R. M.;Okarma, P. J. J. Org.Chem. 1985, 50, 1999.

F430 Model Chemistry H-Ni"(OEiBC)-

+ HA

-

(H),Ni"(OEiBC)

(H),Ni"(OEiBC) H-Ni"(OEiBC)-

+ HA

Inorganic Chemistry, Vol. 31, No. 24, 1992 4999

Ni"(0EiBC)

+ A-

+ H,

(v*-H,)Ni"(OEiBC)

(T2-H2)Ni"(0EiBC)

Ni"(0EiBC)

+ A-

+ H,

(28)

(29) (30)

(31)

tonation on the hydride, eq 30, affords an $-H2 complex, which then dissociates H2, eq 3 1. Dihydrogen complexes of metalloporphyrinsa r e k n ~ w n The . ~ ~ isotopicexchangeof H/D between D2 and EtOH that is catalyzed by a Ni-salicylaldehyde thiosemicarbanzonecation was proposed to involvea nickel dihydrogen intermediate.38 An objection might be raised to the reactions proposed in eqs 25, 27, and 29 and for that matter in eqs 8 and 9. Reductive elimination, @-hydride elimination, and migratory insertion reactions require cis coordination ~ites.2~ Porphyrins and tetraazamacrocyclegenerally are effective at blocking all available coordination sites cis to an axial ligand.25a Thus, the apparent CO insertion into the Fe-C bond of alkyliron(II1) porphyrins to afford an acyliron(II1) p0rphyrin3~or into the Rh-H bond of hydridorhodiwn porphyrins to afford a formylrhodiumporphyrin4 follows radical mechanisms rather than the conventional organometallic mechanisms. Similarly, @-hydrideelimination in the macrocyclic complex R-C~(dmg-BF~)~pyproceeds by transfer of a @-hydrogenof the alkyl group (formally as a hydride) to a macrocycle nitrogen atom.41 The structural characteristics of Ni(1) and the flexibility of isobacteriochlorinsmight obviate the objection. Ni(1) complexes of tetraazamacrocycle~,~~ F430M,43 and i~obacteriochlorins~ have been characterized by EXAFS spectroscopy and, in the case of one of the tetraazamacrocyclic complexes, by X-ray crystallography. In all cases, reduction of Ni(I1) to Ni(1) results not only in an expansion of the macrocycle core but also in a significant distortion in the Ni-N bond lengths within the complex. For the tetrapyrrole complexes, the change is from four equivalent Ni-N distances of roughly 1.90-1.93 A in the low-spin Ni(I1) forms of the complexes to two Ni-N distances of 1.85-1.88 A and two Ni-N distances of 2.00-2.03 A in the Ni(1) forms of the complexes. Preliminary XAFS data for Ni'(0EiBC)- are consistent with these observation^.^^ The distortion of the Ni(1) coordination sphere is unexpected and unexplained. Although Ni(1) is d9 and subject to JahnTeller distortions, the tetragonalligand field associatedwith absent or weakly coordinating axial ligands would give the required nondegenerate ground state. EXAFS data for the isoelectronic d9 Cu(0EiBC) are consistent with four equivalent Cu-N distances of 2.00 A.44 One possibility is that the macrocycle begins to distort toward a trigonal bipyramidal environment around Ni(1). The Ni-N distanceswill split into two distinct sets if two nitrogens each occupy "axial" and "equatorial" sites. On the basis of

(40) (41) (42) (43)

Collman, J. P.; Wagenknecht, P. S.;Hembre, R. T.;Lewis, N. S.J. Am. Chem. Soc. 1990, 112, 1294. Zimmer, M.;Schulte, G.; Luo, X.-L.; Crabtree, R. H. Angew. Chem., Inr. Ed. Engl. 1991, 30, 193. (a) Arafa, I. M.; Shin, K.; Goff, H. M. J. Am. Chem. SOC.1988, 110, 5228. (b) Gueutin, C.; Lexa, D.; Momenteau, M.; Saveant. J.-M. J . Am. Chem. SOC.1990, 112, 1874. (a) Wayland, B. B.;Woods, B. A. J . Chem.Soc., Chem. Commun. 1981, 700. (b) Paoness, R. S.; Thomas, N. C.; Halpern, J. J . Am. Chem. SOC. 1985, 107. 4333. Shi, S.; Bakac, A.; Espenson, J. H. Inorg. Chem. 1991, 30, 3410. Furenlid, L.R.;Renner, M. W.; Szalda, D. J.; Fujita, E. J . Am. Chem. SOC.1991, 113, 883. Furenlid, L. R.; Renner, M. W.; Fajer, J. J . Am. Chem. SOC.1990.11 2,

8987. (44) Fajer, J.; Furenlid, L. R.;Lahiri, G. K.; Renner, M . W.; Stolzenberg, A. M. To be published.

molecular mechanics calculations, Zimmer and Crabtree suggested that F430 can bend to accommodate a trigonal bipyramidal "tacolike" conformation of this type.45 At present, the available evidence suggests that Ni(1) complexes are four-coordinate.g.42-44 The nascent fifth site (an equatorial site) could be occupied by a lone pair in an orbital derived from the original d,z or by a loosely coordinated counterion. Some reduced Ni tetraazamacrocycle complexescoordinate CO to afford true 5-coordinate Ni(1) c0mplexes.~2EXAFS data for these complexes suggest that the Ni-C distance is short and that Ni(1) moves out of the plane of the four nitrogens. These observations are consistent with a trigonal bipyramidal (or square pyramidal) structure. The structures of the R-Ni(0EiBC) complexes should be similar to a first approximation. A sixth, cis coordination site should be relatively easy to open, given that the Ni is already out of plane and the macrocycle is already bent. This would permit the reactions in eqs 8, 9, 25, 27, and 29 to proceed via normal organometallic mechanisms. Implications of Trace Products. Trace quantities of ethylene and butene (lo4 to mol/mol of Ni(1)) are produced during the respective reductions of CH3X and CZHSX,where X = I and OTs. Control experimentsshow that formation of these products depends on the presence of NiL(0EiBC)- and is not artifactual. Any mechanism proposed for the reactions of NiI(0EiBC)- must explain how these products form. It is difficult to envision reactions in which methyl radicals afford ethyleneor ethyl radicals afford butene. However, extensions of the nickel-based reactions proposed above can provide reasonable explanations. Ethylene appears to result from the reaction of NiI(0EiBC)with CH212, which GC analyses show is present as a roughly 0.1?6impurity in some samplesof CHJ. Reaction of Ni'(0EiBC)with pure CH212 affords significant quantities of ethylene and lesser amounts of methane. When mixed, CHJ2 competes successfully with CH31for Nil(OEjBC)-. A close precedent is found in the reaction of K2[Ru(TPP)] with CH2X2 (X = C1, I) in THF to afford a mixture of Ru(TPP)(CHz=CH2) and R U ( T P P ) ( T H F ) ~ .Direct ~ ~ evidence for a Ru=CH2 carbene intermediate was not found. However, bimolecular coupling of M=CH2 complexes to afford mixtures of M(CH2=CH2) and M(so1vent) complexes is known.47 Loss of ethylene from (CHZ=CH2)Ni(OEiBC) should be facile. Obviously, a CH212 impurity cannot account for ethylene formed in reactions of NiI(0EiBC)- with CH30Ts. The ethylidene complex Ru(OEP)(CHCH3) results from Ru(OEP)(CHzCH3)2 when homolysis of the latter affords an ethyl radical which then abstracts an cy-hydrogenfrom a second molecule of the starting complex.48 Production of "Ni(OEiBC)(CH2)" by an analogous mechanism must occur at trace levels if, as argued above, the contribution of free radical chemistry to the overall mechanism is relatively unimportant. The formation of butene from C2H5Xcan be explained by a simple reaction sequence. Significant amounts of ethylene are produced from C2HsX. Therefore, some ethylenewill be present in solution during the reaction. Insertion of ethylene into an ethyl-nickel bond affords butylnickel complexes, eq 32. Butene results from P-hydride elimination, eq 33. An insertion step analogous to eq 32 is proposed in the coupling of alkyl bromides and acrylonitrile that is mediated by the Ni(1) complex Ni(teta)+."b Biological Significance. Ni(0EiBC) and F430 differ in some important respects. The hydrocorphinoid macrocycle of F430 (45) Zimmer, M.; Crabtree, R. H. 1.Am. Chem. SOC.1990, 112, 1062. (46) Collman, J. P.; Brothers, P. J.; McElwee-White, L.; Rose, E. J . Am. Chem. SOC.1985, 107, 61 10. (47) (a) Schrock, R. R. J . Am. Chem. SOC.197597,6577. (b) Merrifield, J. H.; Lin, G.-Y.; Kiel, W. A.; Gladysz, J. A. J . Am. Chem. SOC.1983, 105, 5811. (48) Collman, J. P.; McElwee-White, L.; Brothers, P. J.; Rose, E. J. Am. Chem. SOC.1986, 108, 1332.

Lahiri et al.

5000 Inorganic Chemistry, Vol. 31, No. 24, 1992

CH2=CH2

+ CH,CH,Ni(OEiBC)”

-

CH,CH,CH,CH,Ni(OEiBC)” CH3CH2CH,CH2Ni(OEiBC)” C4H,

-

+ HNi(0EiBC)”

(n = 0 , l ) (32)

(n = 0, 1) (33)

coordinates as a monoanionic ligand. In contrast, OEiBC is a dianion. Thus, the Ni(I1) form of F430M is a monocationic complex whereas Ni(0EiBC) is a neutral complex. Reactions with organometallicreagents like (CHs)2Mgthat are facile with cationic F430M fail to occur for neutral Ni(OEiBC), because the absence of an anionic counterion eliminatesthe large favorable contribution to M t h a t is associated with the formationof MgX2. Furthermore, the positive charge contributes to the roughly 600mV anodic shift of the reduction potential of the Ni(II)/Ni(I) coupleof F430 M relative t0Ni(OEiBC).2,7,8,9a.~~,~~ The presence of the cyclohexanone ring in F430 is also thought to contribute to the anodic shift.98 The shift in reduction potentials appears to make NiI(0EiBC)- considerably more reactive than NilF430M. Despite these differences, the reactivities of Ni’(0EiBC)- and of Ni’F430M share many common features. Both react rapidly with CH31and more slowly with CH30Tsto afford CH4 and the Ni(I1) complex. The fact that NiIF430M reacts with CH30Ts suggests strongly that Ni1F430M, like Ni’(OEiBC)-, reacts by a two-electron mechanism rather than by a single-electron transfer mechanism. Labeling studies show that the fourth hydrogen in methane is transferred in a protonation step, presumably of an alkylnickel intermediate, and appears to be derived from residual water in the solvent. Compelling evidence exists for the methylnickel F430 intermediate. Addition of CH31to NiIF430M at -60 OC causes an immediate color change, but CH4 evolution only occurs slowly on warming to room temperature or immediately upon addition of 0.1 M aqueous HCl.Io 2H NMR spectroscopy provides direct evidence for the existence of CD3nickel F430M, prepared by reaction of F430M and (CD&Mg at -78 O c a z 7 Evidence for the intermediacy of R-Ni(0EiBC) complexes is circumstantial, because the complexes appear to be much more reactive. There is no perceptible lag betweenoxidation and CHI evolution upon low-temperature addition of CH31 to NiI(OEiBC)-. The coupling product ethane is not mentioned as a product of thereaction ofNiIF430M with CH31in DMF.IO However, ethane is a product of the F430-mediated reduction of polychlorinated methanes by titanium(II1) citrate in aqueous s ~ l u t i o n . ~In addition to methane and ethane, the second reaction also produces significant quantities of nonvolatile, yet unidentified, products. Thus, the yield of volatile hydrocarbon products is less than quantitative, just as we observe for NiI(OEiBC)-in some solvents. The most curious similarity in the reactivities of NiI(0EiBC)and of the Ni(1) form of F430 is that neither appears to react with thioethers. This includes methyl coenzyme M, CH3SCH2CH2SO3-, the natural substrate of F430 containing methyl coenzyme M reductase. Does this imply that the reactivity of

Nil(OEiBC)-and of Ni’F430 is irrelevant to the chemistry carried out by the enzyme? It is premature to draw such a conclusion. The reactivity of the Ni(1) complexes is almost certainly of direct relevance to the extent that the enzyme is involved in the reductive dehalogenation of chlorinated hydrocarbons by methanogenic bacteria. The situation is less certain as regards enzymatic catalysis of eq 1. Several possibilities remain open. First, the enzymatic mechanism could involve chemistry based on Ni(II1) species rather than on Ni(1) species. Second, the failure of Ni(1) complexes to react directly with methyl coenzyme M might be apparent rather than real. Other more rapid reactions could consume Ni(1) before it had time to react with a thioether. Reactions of Ni’F430M with proton donors have not been investigated. Methyl coenzyme M was almost certainly added as the ammonium salt. Did the ammonium ion rapidly oxidize Ni1F430Mand form hydrogen as in the case of Ni’(OEiBC)-? Third, methyl coenzyme M could require activation prior to reaction with Ni’F430.10 One proposal is that the thiyl radical of cofactor HTP, which is postulated to result from one-electron reduction of F430 by HTP-S-, couples with methyl coenzyme M to afford the sulfuranyl radical, HTPSS*(CH3)(CoM).49~So The sulfuranyl radical should be a better alkylating agent than a thioether. Finally, Ni’F430 could require activation by conversiontoa formother than present inNi’F430M. Some possibilities include coordination of a protein side chain, coordination of the thio cofactor HTPS-H, or protonation by HTPS-H. The latter would imply that a nickel hydride is the reactive intermediate rather than a Ni(1) complex. In essence, the enzyme would function as a biological equivalent of Raney nickel. Formation of a small amount of methane in the reaction of NiI(0EiBC)- with the ammonium salt of methyl CoM is interesting in this light. It must be stressed that the failure to observe methane when tetraphenylphosphonium methyl CoM and thiophenol were simultaneouslyadded to NiI(0EiBC)- does not eliminate the hydride intermediate from consideration. The rate of oxidation of NP(0EiBC)- by the acidic thiophenol may be too rapid to permit any nickel hydride to react with methyl CoM. A significant conclusion can be drawn from our results if NiIF430 is involved in the chemistry of methyl coenzyme M reductase. The protein environment of the enzyme is extremely important in imposing one reaction pathway on a very reactive reduced nickel species.

Acknowledgment. This work was supported by the National Institutes of Health (Grant GM 33882). Supplementary Material Available: Text containing discussions of construction and operation of the gas sample valve system, calibration of GC detectors, independent confirmation of analytical methods, and control experiments and a figure showing a line drawing of the gas sample valve system (9 pages). Ordering information is given on any current masthead page.

B. Helv. Chim. Acta 1990, 73, 2209. (50) Berkessel, A. Bioorg. Chem. 1991. 19, 101. (49) Jam,